Shopping trolley AF sensor “holds promise for stroke prevention”


The use of embedded electrocardiogram (ECG) sensors in shopping trolley handles could effectively identify individuals with previously undiagnosed atrial fibrillation (AF). This is among the findings from the SHOPS-AF study, presented at the annual congress of the Association of Cardiovascular Nursing and Allied Professionals (ACNAP 2023, 23–24 June, Edinburgh, Scotland).

“This study shows the potential of taking health checks to the masses without disrupting daily routines,” said study author Ian Jones (Liverpool John Moores University, Liverpool, UK). “Over the course of two months, we identified 39 patients who were unaware that they had AF. That’s 39 people at greater risk of stroke who received a cardiologist appointment.”

Ten trolleys had a sensor placed in the handle and were used across four supermarkets with pharmacies in Liverpool during the two month study. Shoppers were asked to use a modified trolley and hold the handlebar for at least 60 seconds.

If the sensor did not detect an irregular heartbeat, it lit up green. These participants had a manual pulse check by a researcher to confirm there was no AF. If an irregular heartbeat was found, the sensor lit up red. The in-store pharmacist then did a manual pulse check and another sensor reading using a standalone bar not attached to a trolley with the participant standing still.

The study cardiologist reviewed the ECG recordings of participants with a red light and/or irregular pulse. Participants were informed of the results, which were: 1) no AF; 2) unclear ECG and an invitation to repeat the measurement; or 3) AF confirmed and a cardiologist appointment within two weeks.

A total of 2,155 adults used a shopping trolley. ECG data were available for 220 participants who either had a red light on the sensor and/or an irregular pulse, suggesting AF. After ECG review by the study cardiologist, there was no evidence of AF in 115 participants, 46 recordings were unclear, and AF was diagnosed in 59 participants. The average age of the 59 participants with AF was 74 years and 43% were women. Of those, 20 already knew they had AF and 39 were previously undiagnosed.

To assess the accuracy of screening using this method, the researchers conducted three analyses: 1) excluding all 46 unclear ECGs; 2) assuming all unclear ECGs were AF; and 3) assuming all unclear ECGs were not AF. This showed that the sensor’s sensitivity ranged from 0.70 to 0.93 and specificity ranged from 0.15 to 0.97. This resulted in a positive predictive value of 0.24 to 0.56, meaning that only one-quarter to one-half of those found to have AF according to the sensor and/or manual pulse check actually had the condition (i.e. there were a high number of false positives). The negative predictive value was 0.55 to 1.00, meaning that around half of actual atrial fibrillation cases would be missed using this method (i.e. false negatives).

Jones said: “Nearly two-thirds of the shoppers we approached were happy to use a trolley, and the vast majority of those who declined were in a rush rather than wary of being monitored. This shows that the concept is acceptable to most people and worth testing in a larger study. Before we conduct SHOPS-AF II, some adjustments are needed to make the system more accurate. For example, having a designated position on the bar to hold onto, as hand movement interfered with the readings. In addition, European Society of Cardiology (ESC) guidelines require just a 30 second ECG to diagnose AF, so we aim to find a sensor that will halve the time shoppers need to continuously hold the bar.”

He concluded: “Checking for AF while people do their regular shopping holds promise for preventing strokes and saving lives. A crucial aspect is providing immediate access to a health professional who can explain the findings and refer patients on for confirmatory tests and medication if needed.”


Please enter your comment!
Please enter your name here